After unsuccessful attempts to determine the origin of the dynamical stability of cubic NbN using Density Functional Perturbation Theory (DFPT) in Quantum ESPRESSO, we shifted our approach to machine-learned interatomic potentials (MLIP). This method allows us to investigate the effects of vacancy concentration and atomic site distribution on dynamical stability at different temperatures. To implement this approach, I first needed to learn the methodology, which led to an idle period in my activity. Currently, the MLIP models are in training. Once trained, we will use them to study phonon dispersion across different temperatures, ultimately identifying the origin of stability in cubic NbN.